Image Recognition Method for Micropores Inside Small Gas Pipelines

Round 1

Reviewer 1 Report

The manuscript presents a promising approach for addressing the challenge of detecting micro pores in small gas pipelines using image recognition. The comprehensive network design and empirical validation enhance the significance of the research. With some minor improvements in detailing the methodology and discussing the practical implications of the results, it well poised the manuscript for acceptance.

 

Suggestions for Improvement:

1. Clarity of Methodology: While the abstract provides a general overview of the methodology, it would be beneficial to include a bit more detail on the specific techniques or algorithms used within the feature fusion network and classification prediction network.
2.  
3. Dataset Description: It would be helpful to briefly mention the nature and size of the micro pores datasets used for training and validation. This could provide insights into the diversity and representativeness of the data.
4.  
5. Discussion of Results: Consider expanding the abstract to briefly discuss the implications of achieving a precision of 94.7%, a detection rate of 96.6%, and an average precision of 95.5% for the practical application of detecting small gas pipeline leaks.

Author Response

请参阅附件。

Author Response File: Author Response.docx

Reviewer 2 Report

The topic of the paper is very interesting. However, the paper is not well-organized. The contributions are not clear. The authors should specify whether they have collected their own images (from the robot) or if they took a ready dataset. The paper is poorly written and the flow is not clear. The authors should consider the following comments.

1.      The authors should not write the paper as bullets. Instead, it should be a concise text with clear paragraphs and sections.

2.      The YOLO5 algorithm is well-known in image processing. The authors didn’t show what are the novelties of their study compared to previous studies using the same algorithm.

3.      In subsection 4.1.2, the authors should clarify how did they choose the algorithm hyperparameters.

4.      The manuscript is missing a conclusion section.

5.      The contributions of the paper are not clearly presented (they should be provided at the end of the introduction section).

6.      The paper organization is poor and the figures’ quality is also poor.

 

 

 

English is poor and should be revised. 

Author Response

请参阅附件。

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The paper can be accepted in its current form. 

English should be polished more.